Whitespace needed between double arguments? (german locale)

[mlarue]

class Math {
    static add(a, b) {
        return a + b
    }
}

var result = Math.add(10,13)
System.print("add = %(result)")

if (result != 23) {
    Fiber.abort("result should be 23, but is: %(result)")
}

When I load this script with wrenInterpret it fails with: result should be 23, but is: 23,13

It works, if I add a whitespace between the arguments like so: Math.add(10, 13)

If i run this script with bin/wren_test everything works as expected. Both with and without the whitespace.

[mhermier]

Smells like a bug in your evaluation of the string code. But I have no idea why it would do that...

[mlarue]

Actually I'm playing around with java-ffm and the problem seems to be there.

I added my script to example/embedding/main.c and it works as expected.

So probably NOT a bug here..

#include <stdio.h>
#include "wren.h"

static void writeFn(WrenVM* vm, const char* text)
{
  printf("%s", text);
}

void errorFn(WrenVM* vm, WrenErrorType errorType,
             const char* module, const int line,
             const char* msg)
{
  switch (errorType)
  {
    case WREN_ERROR_COMPILE:
    {
      printf("[%s line %d] [Error] %s\n", module, line, msg);
    } break;
    case WREN_ERROR_STACK_TRACE:
    {
      printf("[%s line %d] in %s\n", module, line, msg);
    } break;
    case WREN_ERROR_RUNTIME:
    {
      printf("[Runtime Error] %s\n", msg);
    } break;
  }
}

int main()
{

  WrenConfiguration config;
  wrenInitConfiguration(&config);
    config.writeFn = &writeFn;
    config.errorFn = &errorFn;
  WrenVM* vm = wrenNewVM(&config);

  const char* module = "main";
  //const char* script = "System.print(\"I am running in a VM!\")";
  const char* script =
    "class Math {\n"
    "    static add(a, b) {\n"
    "        return a + b\n"
    "    }\n"
    "}\n"
    "\n"
    "var result = Math.add(10,13)\n"
    "System.print(\"add = %(result)\")\n"
    "\n"
    "if (result != 23) {\n"
    "    Fiber.abort(\"result should be 23, but is: %(result)\")\n"
    "}\n";

  WrenInterpretResult result = wrenInterpret(vm, module, script);

  switch (result)
  {
    case WREN_RESULT_COMPILE_ERROR:
      { printf("Compile Error!\n"); } break;
    case WREN_RESULT_RUNTIME_ERROR:
      { printf("Runtime Error!\n"); } break;
    case WREN_RESULT_SUCCESS:
      { printf("Success!\n"); } break;
  }

  wrenFreeVM(vm);

}

gcc -I../../src/include main.c -L../../lib -l:libwren.a -lm -o main

[ruby0x1]

Make sure the string is null terminated, that's a common cause of confusing errors when interfacing around

[mlarue]

I paste my java method for completeness here. Should anybody be curious.. ;)

    @Test
    public void basicClassMethod() {
        Arena arena = Arena.ofShared();

        MemorySegment configuration = arena.allocate(WrenConfiguration.layout());
        wren_h.wrenInitConfiguration(configuration);

        WrenConfiguration.writeFn(configuration, WrenWriteFn.allocate((MemorySegment vm, MemorySegment text) -> {
            System.out.print(text.getString(0));
        }, arena));

        WrenConfiguration.errorFn(configuration, WrenErrorFn.allocate(
                (MemorySegment vm, int type, MemorySegment module, int line, MemorySegment message) -> {
                    String moduleName = module.address() != wren_h.NULL().address() ? module.getString(0) : "(unknown)";
                    System.err.println("module=" + moduleName + ", line=" + line + ", type=" + type +
                                       ", message=" + message.getString(0));
                }, arena));

        MemorySegment vm = wren_h.wrenNewVM(configuration);

        int result = wren_h.wrenInterpret(
                vm,
                arena.allocateFrom("my_module"),
                arena.allocateFrom("""
                        class Math {
                            static add(a, b) {
                                System.print("a = %(a)")
                                System.print("b = %(b)")
                                return a + b
                            }
                        }
                        
                        var result = Math.add(10,13)
                        System.print("add = %(result)")
                        
                        if (result != 23) {
                            Fiber.abort("result should be 23, but is: %(result)")
                        }
                        """));

        System.out.println("result: " + result);
    }

Output:

a = 10,13
b = 13
add = 23,13
module=(unknown), line=-1, type=1, message=result should be 23, but is: 23,13
module=my_module, line=13, type=2, message=(script)
result: 2

[mhermier]

My guess is that it is a locale issue. Invocation of strtod accepted 10,13 as a number. That also explains why 10, 13 works correctly

[mlarue]

@mhermier

That's it! Thanks.

This works:

export "LC_ALL=C" && mvn test -Dtest=LowLevelTest#basicClassMethod

See also:

• https://stackoverflow.com/questions/78717017/java-foreign-function-memory-api-problems-with-locale

[mlarue]

I will test the C version again with German locale and reopen the issue if necessary.

[mlarue]

Ok. As expected the C version also has trouble parsing doubles with German locale:

setlocale(LC_ALL, "de_DE");

...

  const char* script = "class Math {                         \n"
                        "    static add(a, b) {              \n"
                        "        System.print(\"a = %(a)\")  \n"
                        "        System.print(\"b = %(b)\")  \n"
                        "        return a + b                \n"
                        "    }                               \n"
                        "}                                   \n"
                        "                                    \n"
                        "var result = Math.add(10,13)        \n"
                        "System.print(\"add = %(result)\")   \n"
                        "                                    \n"
                        "if (result != 23) {                 \n"
                        "    Fiber.abort(\"result should be 23, but is: %(result)\")\n"
                        "}                                   \n";

  WrenInterpretResult result = wrenInterpret(vm, module, script);

Output:

a = 10,13
b = 13
add = 23,13
[Runtime Error] result should be 23, but is: 23,13
[main line 13] in (script)
Runtime Error!

Setting a to "10,13" and b to "13" looks like a bug.

I think a possible outcome could be: "2 arguments expected, but got only 1"

But Math.add(10,13, 100) also does not look correct (and does not work). Maybe "strtod" respecting locales is not helpful here. The script parser should always use something like LC_ALL=C?

[mlarue]

My workaround: I do not change the locale for the whole process (#1214 (comment)), but only for the wrenInterpret call:

+WrenInterpretResult wrenInterpretWithCLocale(WrenVM* vm, const char* module,
+                                             const char* source)
+{
+  locale_t loc = newlocale(LC_ALL_MASK, "C", (locale_t)0);
+  locale_t original = uselocale(loc);
+  WrenInterpretResult result = wrenInterpret(vm, module, source);
+  uselocale(original);
+  return result;
+}

[mhermier]

A less portable way, but less intrusive, would be to use strtod_l. It is quite portable today, bu not C99 compliant. Le lun. 5 janv. 2026, 12:43, Markus Lauer ***@***.***> a écrit :

…

*mlarue* left a comment (wren-lang/wren#1214) <#1214 (comment)> My workaround: I do not change the locale for the whole process (#1214 (comment) <#1214 (comment)>), but only for the wrenInterpret call: +WrenInterpretResult wrenInterpretWithCLocale(WrenVM* vm, const char* module, + const char* source) +{ + locale_t loc = newlocale(LC_ALL_MASK, "C", (locale_t)0); + locale_t original = uselocale(loc); + WrenInterpretResult result = wrenInterpret(vm, module, source); + uselocale(original); + return result; +} — Reply to this email directly, view it on GitHub <#1214 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AGHNXFO4BEUNJBY35DSTOIL4FJE57AVCNFSM6AAAAACQSWROXOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTOMJQGA4TGMBTGQ> . You are receiving this because you were mentioned.Message ID: ***@***.***>